Xen and the art of virtualization
SOSP '03 Proceedings of the nineteenth ACM symposium on Operating systems principles
User-Driven Scheduling of Interactive Virtual Machines
GRID '04 Proceedings of the 5th IEEE/ACM International Workshop on Grid Computing
Xen and co.: communication-aware CPU scheduling for consolidated xen-based hosting platforms
Proceedings of the 3rd international conference on Virtual execution environments
Comparison of the three CPU schedulers in Xen
ACM SIGMETRICS Performance Evaluation Review
Scheduling I/O in virtual machine monitors
Proceedings of the fourth ACM SIGPLAN/SIGOPS international conference on Virtual execution environments
The definitive guide to the xen hypervisor
The definitive guide to the xen hypervisor
Task-aware virtual machine scheduling for I/O performance.
Proceedings of the 2009 ACM SIGPLAN/SIGOPS international conference on Virtual execution environments
The hybrid scheduling framework for virtual machine systems
Proceedings of the 2009 ACM SIGPLAN/SIGOPS international conference on Virtual execution environments
Enforcing performance isolation across virtual machines in Xen
Proceedings of the ACM/IFIP/USENIX 2006 International Conference on Middleware
VM3: Measuring, modeling and managing VM shared resources
Computer Networks: The International Journal of Computer and Telecommunications Networking
ICPADS '09 Proceedings of the 2009 15th International Conference on Parallel and Distributed Systems
Survey of virtual machine research
Computer
The Journal of Supercomputing
| Hi-index | 0.00 |
Due to the advances in software and hardware support for virtualisation, virtualisation technology has been adapted for server consolidation and desktop virtualisation to save on capital and operating costs. The basic abstraction layer of software that virtualises hardware resources and manages the execution of virtual machines is called virtual machine monitor (VMM). A critical part of VMM is the CPU scheduler which slices and dispatches physical CPU time to virtual machines. Xen's credit scheduler utilised blocked-to-boosted mechanism to achieve low latency on I/O intensive tasks. However, it suppresses event notifications for the guest domain that is not blocked. This may delays the response of a guest domain doing mixed workloads, as its virtual CPU is seldom blocked when processing CPU-intensive tasks. We enhance the credit scheduler by making it full-time aware of inter-domain events and physical interrupt request events. Our proposed scheduler not only improves the responsiveness of domains doing mixed workloads, but also minimises the possibly caused scheduling unfairness. The experimental evaluation demonstrates the benefits of our proposed scheduler.